Lundahl themselves indicate to have a better transformer for a 300B, and they are right.
Using an LL9202 with 6k5 primary load you either have:
1. a 600 ohm primary DC resistance and 100H of primary inductance, or
2. a 150 ohm primary DC resistance and 25H of primary inductance.
Both options are rather sub optimal for a 300B amplifier.
There is a third option to connect the LL9202 to better suit a 300B:
- connect three primaries in series, and connect the fourth primary parallel with one of the three;
- create two strings of secondaries by series connecting four of them, and parallel the two.
Now you have a 3k16 primary load with 8 ohm secondary.
Actually a bit higher primary load because of the highish primary DC resistance which will be a "bad" 375 ohm.
But once more: quite sub optimal....
Andy, I don't think that the WE datasheet is representative for all 300B's.
Using an LL9202 with 6k5 primary load you either have:
1. a 600 ohm primary DC resistance and 100H of primary inductance, or
2. a 150 ohm primary DC resistance and 25H of primary inductance.
Both options are rather sub optimal for a 300B amplifier.
There is a third option to connect the LL9202 to better suit a 300B:
- connect three primaries in series, and connect the fourth primary parallel with one of the three;
- create two strings of secondaries by series connecting four of them, and parallel the two.
Now you have a 3k16 primary load with 8 ohm secondary.
Actually a bit higher primary load because of the highish primary DC resistance which will be a "bad" 375 ohm.
But once more: quite sub optimal....
Andy, I don't think that the WE datasheet is representative for all 300B's.
Last edited:
As I said, I compared LL1664/70 directly with LL1682/50 on a few occasions with my 300b SE and it was no contest - the LL1682 clearly sounded better, and it wasn't subtle. Speakers were Alpair 10M. The LL1682/50 would not be considered an obvious match for a 300b at 5K into 5R. You might even call it "sub optimal" but in practice it works audibly better. I'm not the only 300b user that prefers a 5K OPT, and it's a small step to go a little higher. I was using more like 320v a-k. You don't have to up as high as 450v. Depends on how much power you need. WE data attached. Distortion is low at 350v a-k and 50mA into 5K.
I was using EH 300b tubes. Could be differences between makes - I can't comment on that.
.
Attachments
Last edited:
Hey SS - the 25T did come to mind as it is on my short list to build with, another A2 design with NFB. 3.25K 100mA OPT would suit well I think if the transformer is rewired. I have a batch of 25T (3C24) with Eimac grid / plate caps on hand. It biases with a positive grid voltage, had thought I might be able to forgo a negative supply if I choke load a direct-coupled cathode follower for the negative grid swings. Need to generate enough gain on the input for the feedback loop, but with the high mu of the 25T, very little closed loop gain will be needed for a good input sensitivity. Will keep this one in mind, might even try breadboarding it, I have a suitable mains transformer as well.
Would love to build with the type 50, but the price is intimidating! I was very seriously considering it at one point, even reached out to EML who say they are continuing production of their type 50. But seeing as a type 50 SET is a cost-no-object design already, I think I would want to get dedicated transformers. Will keep it in mind though 😉
Another good idea, thanks Andy, my speakers are 92dB/W, so do not need a large power output for my listening space, I am happily running them at 3.75W at the moment, so the higher primary Z with lower distortion and a slight power loss is no problem I think. Will think this over as well, interesting results with your 5K transformer.
Are you basing this on actual experience of listening to Lundahl OPTs at 2.5K and 5K? Or is this an abstract conclusion? Also, what operating points are you recommending here?
Well TJ full music has type 50s for ~300 bucks a pair
50 (TJ Fullmusic. Mesh Plate. Matched Pair.) – DIY Hifi Supply
Someone also has a collection of NOS national unions on ebay for 500 bucks a pair. You don't have to pay 1000 bucks to EML to get a type 50 : )
While I am on the topic of TJ full music, take a look at the PX4 and the PX25. Both of those would definitely make an interesting amp as well.
50 (TJ Fullmusic. Mesh Plate. Matched Pair.) – DIY Hifi Supply
Someone also has a collection of NOS national unions on ebay for 500 bucks a pair. You don't have to pay 1000 bucks to EML to get a type 50 : )
While I am on the topic of TJ full music, take a look at the PX4 and the PX25. Both of those would definitely make an interesting amp as well.
2.5K has traditionally been used for good reasons. A smaller primary inductance gives lower distortion, deeper bass response and a less hard and thin sound.
N101N,
I find your Post # 27 somewhat mysterious.
So just parallel a 5H or 10H choke in parallel with the 2.5k, 3.5k, or 5k primary.
Lower distortion?
Deeper bass response?
Less hard and thin sound?
How does that work?
Or, just use less Primary turns and less Secondary turns (keep the proper P:S ratio for 2.5k primary).
Lower distortion?
Deeper bass response?
Less hard and thin sound?
How does that work?
Please explain the theory behind the observations.
I find your Post # 27 somewhat mysterious.
So just parallel a 5H or 10H choke in parallel with the 2.5k, 3.5k, or 5k primary.
Lower distortion?
Deeper bass response?
Less hard and thin sound?
How does that work?
Or, just use less Primary turns and less Secondary turns (keep the proper P:S ratio for 2.5k primary).
Lower distortion?
Deeper bass response?
Less hard and thin sound?
How does that work?
Please explain the theory behind the observations.
Last edited:
There is no tradition with 300b - anything from 2.5K up to 5K is common. But we're looking at 2 factors when it comes to distortion - the gap and the primary value. For instance Lundahl quote the same inductance for LL1664 at 3K, LL1663 at 5K and LL1682 at 5K into 5R. It's 35H for the 50mA gap and 17H for the 100mA gap. My preference for LL1682/50 over LL1664/70 may be partly due to the higher primary value and partly the higher inductance. And then we have the operating point as another variable, the speaker load as another. In my case the speaker and operating point were the same. My experience with OPTs and plate chokes has been that a higher inductance gives a cleaner sound. With plate chokes, for instance, I've found on two occasions that putting 2 plate chokes in series sounded better than just one. Your ears and taste may be different from mine, of course.
Simply put, that is due to the larger current (mass). Inductive reactance is the key parameter here.
@ Andy, 6A3sUMMER,
You're feeding a troll...
At least, based on his posts and blanket statements, that's my opinion.
I would not pay him attention.
You're feeding a troll...
At least, based on his posts and blanket statements, that's my opinion.
I would not pay him attention.
Of course, for those of you who took my post the wrong way . . .
I did not agree with N101N's post # 27 statements.
That is why I gave a couple of non-sensical "solutions".
That is why I asked him to explain what he said.
If the combination of the current through the primary, primary turns, the air gap spacing, and the lamination size are not correct, there will be early saturation of the core.
That causes distortion.
It can also cause the low frequencies to start rolling off at a higher than normal frequency.
As always, the loudspeaker load versus frequency, DCR, pure resistive impedance, and reactive impedance, affects the amplifier performance.
For an example of optimizing the match . . .
Take a look at Eduardo B DeLimas article in Valve magazine (A VSAC and Doc BottleHead publication).
Eduardo, are you still out there?
I did not agree with N101N's post # 27 statements.
That is why I gave a couple of non-sensical "solutions".
That is why I asked him to explain what he said.
If the combination of the current through the primary, primary turns, the air gap spacing, and the lamination size are not correct, there will be early saturation of the core.
That causes distortion.
It can also cause the low frequencies to start rolling off at a higher than normal frequency.
As always, the loudspeaker load versus frequency, DCR, pure resistive impedance, and reactive impedance, affects the amplifier performance.
For an example of optimizing the match . . .
Take a look at Eduardo B DeLimas article in Valve magazine (A VSAC and Doc BottleHead publication).
Eduardo, are you still out there?
Last edited:
I make some mistakes in my posts.
I am not perfect.
But when I see a post that might steer someone the wrong way (a newbie or less knowledgeable reader) I often jump in and try and counteract those statements with the correct principals and statements.
I am not perfect.
But when I see a post that might steer someone the wrong way (a newbie or less knowledgeable reader) I often jump in and try and counteract those statements with the correct principals and statements.
LL9202 – Bartola(R) Valves
I was just looking at this - Ale really liked the 814 tube at one point. That's another idea. Not a common tube, but there must be some about including a few on eBay for reasonable money.
I was just looking at this - Ale really liked the 814 tube at one point. That's another idea. Not a common tube, but there must be some about including a few on eBay for reasonable money.
andyjevans,
Agreed,
For an SE parafeed amp, putting 2 chokes in series, can improve the low frequency performance.
For my non-sense idea of putting a choke in parallel with the output transformer primary, I meant an SE output transformer that had plate current through it, and now the DC current is divided between the choke and the primary.
So, a primary impedance of 10H in parallel with a 10H choke would make the low frequency roll off 2 times earlier. Not a good idea.
Agreed,
For an SE parafeed amp, putting 2 chokes in series, can improve the low frequency performance.
For my non-sense idea of putting a choke in parallel with the output transformer primary, I meant an SE output transformer that had plate current through it, and now the DC current is divided between the choke and the primary.
So, a primary impedance of 10H in parallel with a 10H choke would make the low frequency roll off 2 times earlier. Not a good idea.
Andy,
I previously said that I would opt for the high quality 3K Lundahl rather than a mediocre 2.5K transformer. Without hesitation.
And I would opt for the 2.5K Lundahl rather than a 3K Lundahl.
I previously said that I would opt for the high quality 3K Lundahl rather than a mediocre 2.5K transformer. Without hesitation.
And I would opt for the 2.5K Lundahl rather than a 3K Lundahl.
Last edited:
